CN104997508A - Automatic decomposition method of array type sEMG (surface EMG) signal - Google Patents
Automatic decomposition method of array type sEMG (surface EMG) signal Download PDFInfo
- Publication number
- CN104997508A CN104997508A CN201510512579.4A CN201510512579A CN104997508A CN 104997508 A CN104997508 A CN 104997508A CN 201510512579 A CN201510512579 A CN 201510512579A CN 104997508 A CN104997508 A CN 104997508A
- Authority
- CN
- China
- Prior art keywords
- semg
- signal
- array
- waveform
- moment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/389—Electromyography [EMG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7221—Determining signal validity, reliability or quality
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7246—Details of waveform analysis using correlation, e.g. template matching or determination of similarity
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
Abstract
The invention provides an automatic decomposition method of an array type sEMG (surface EMG) signal. The automatic decomposition method comprises the following steps: firstly, pretreating the array type sEMG signal, and calculating an initial distribution sequence vector; determining the number of kinematic units, namely, the cluster number, by using time domain subtraction; according to the cluster number, clustering array type sEMG waveforms by using a minimum distance classifier; finally, recalculating a new kinematic unit distribution sequence vector, circulating the executive routine till decomposition is finished, classifying all the distribution sequences, and optimizing the result. The decomposition method is high in accuracy, quick in calculation speed and easy to operate.
Description
Technical field
The present invention relates to a kind of array sEMG signal automatic classifying method.
Background technology
Surface electromyogram signal (surface EMG, sEMG) be utilize surface electrode to detect electromyographic signal from body surface, compare with needle electrode electromyographic signal (Needle EMG, NEMG), it has feature non-invasive, easy to patients, therefore has a extensive future.Experiment shows, utilizes array sEMG can improve the verification and measurement ratio of moving cell (MU), particularly improves detecting and recognition effect of small magnitude moving cell active electrical potential (MUAP).The origin of sEMG is MUAP, and active electrical potential discharged by each motor unit activated in given muscle contraction.Given raise pattern at any one, numerous motor units is activated with asynchronous pattern, and the summation of these motor unit activities constitutes the intensity of electromyographic signal.Array sEMG signal is in fact the summation of multiple motor unit active electrical potential, and its wave amplitude is typically between 1 ~ 5000uv, and frequency range is 10 ~ 400Hz.Clinically, nervimuscular functional status can be understood more all sidedly by array sEMG, differentiate neurogenic and muscle-derived disease, judge the position of nerve injury, degree and recovery, and the detection analysis of array sEMG signal also has to rehabilitation medicine and sports medical science significant.
Array sEMG decomposition is in fact classify to the moving cell granting sequence that sEMG comprises, and extracts each current potential sequence of moving cell.At present, sEMG sorting technique mainly contains: K means clustering algorithm, template matching method, artificial neural network (ANN) algorithm, real time linear aliasing blind signal separation algorithm, independent element divide the methods such as folding (ICA), convolution kernel backoff algorithm.K-means clustering algorithm needs the classification number of specifying cluster, and the priori of the unit granting that lacks exercise in electromyographic signal, be difficult to specify accurately classification.Template matching method obtains difficulty due to template, applies limited.ANN can solve containing more superposed waveform situations and eliminate absolute error better when low signal-to-noise ratio, but ANN method is once after training, network just immobilizes, when the shape of template changes, neutral net also needs re-training, so robustness is bad.ICA is a kind of fanaticism decomposition technique, and each MUAPT that its hypothesis forms electromyographic signal is separate, then signal decomposition is become some separate compositions.Convolution kernel backoff algorithm method is a kind of blind signal decomposition method, and the method effect when addition of waveforms is serious is still undesirable.The signal to noise ratio of array sEMG is lower, and the superposition degree that the variability of MUAP waveform is strong and mutual is comparatively large, and this causes its main cause of decomposing difficulty.The decomposition method of plug-in type electromyographic signal is made some amendments and is applied in the disaggregate approach of surface electromyogram signal by Many researchers, but all can not reach the discomposing effect of plug-in type electromyographic signal.As a whole, Decomposition Surface EMG research is also in the exploratory stage, is one of difficult point of myoelectricity research field.
Summary of the invention
In view of the above problems, the object of the present invention is to provide a kind of array sEMG signal automatic classifying method.
For achieving the above object, for array surface electrode electromyographic signal, clustering method based on minimum range device is proposed, granting waveform is classified, and in this process, not solution matrix, obtains the number of muscular movement unit by convolution kernel compensation method and provides sequence, realizing the decomposition of array sEMG.Because the method is considered provide waveform and provide the moment simultaneously, relative to other method, the method has the high advantage of sEMG Decomposition Accuracy.
The invention discloses a kind of array sEMG signal automatic classifying method.It comprises the following steps:
Step one: array sEMG Signal Pretreatment: to sEMG signal filtering, rejects interference;
Step 2: calculate and initially provide sequence vector: the dependency utilizing each channel signal of array sEMG signal, calculates and initially provides sequence vector, as extraction initial value;
Step 3: determine moving cell number: utilize time domain subtraction, obtains moving cell number from sEMG, as cluster number;
Step 4: be poised for battle column sEMG waveform clustering: according to described cluster number, utilize minimum distance classifier, is poised for battle column sEMG waveform clustering;
Step 5: calculate new moving cell and provide sequence vector: according to described waveform clustering result, get the class containing maximum moment, calculate such moment waveform average, calculate new granting sequence vector;
Step 6: to all granting sequence classified finishings: repeat step 2-----step 5, rejects that repeat and irrational granting sequence vector, optimum results.
The technical measures optimized comprise:
The cross-correlation matrix of above-mentioned sEMG signal is expressed as:
Wherein
sampling instant,
?
the array signal of individual sampling instant,
?
the array signal transposition of individual sampling instant,
that number sequence is expected.
Moving cell is provided sequence and is expressed as:
Wherein
the inverse matrix of array signal cross-correlation matrix.
Above-mentioned Euclidean distance formula is expressed as:
Wherein
the value of array sEMG corresponding sometime,
it is the meansigma methods of array sEMG value corresponding to K moment.
The computing formula that new moving cell provides sequence is expressed as:
Wherein
it is N number of wave-average filtering value transposition.
Compared with prior art, a kind of array sEMG signal automatic classifying method of the present invention, when calculating initial granting sequence, because peak-peak is often because interference causes, it is caused to be worth abnormal height, so the present invention adopts is moment corresponding to second largest peak value, obtain providing sequence initial value more accurately; The present invention utilizes the acquisition of time domain subtraction to subtract each other number of times M, and this number of times M, as classification number, in fact provides moving cell number, for the classification of follow-up minimum range device provides priori, improves decomposition accuracy; The present invention utilizes Euclidean distance as classification foundation, realizes convenient, simple; Due to the present invention classification for be waveform, and cross-correlation matrix calculate be provide the moment, so catabolic process consider simultaneously provide the moment and provide waveform, substantially increase the accuracy of decomposition; The present invention does not need to calculate moving cell and provides hybrid matrix between sequence and array sEMG signal, greatly reduces computation time, improves efficiency.This decomposition method running does not need manual intervention, easy to use.
Accompanying drawing explanation
fig. 1 is flow chart of the present invention.
Detailed description of the invention
Be described in further detail the present invention below in conjunction with attached Example, those skilled in the art can the content disclosed by this description realize easily.
Be illustrated in figure 1 flow chart of the present invention.
Step one: array sEMG Signal Pretreatment: to sEMG signal filtering, rejects interference.Owing to comprising various interfering signal in sEMG signal, first pretreatment needs to adopt band filter, retains 10Hz--500Hz frequency band signals, then adopts notch filter, filtering 50Hz Hz noise.
Step 2: calculate and initially provide sequence vector: the dependency utilizing each channel signal of array sEMG signal, calculates and initially provides sequence vector, as extraction initial value.Detailed process is: first computing array sEMG signal cross-correlation matrix and cross-correlation matrix inverse matrix, and cross-correlation matrix is expressed as:
Wherein
sampling instant,
?
the array signal of individual sampling instant,
?
the array signal transposition of individual sampling instant,
that number sequence is expected.Calculate the inverse matrix of cross-correlation matrix
, namely
Then sampling instant
get the intermediate value of sEMG signal energy, energy calculates according to the following formula:
Get the moment corresponding to energy intermediate value
.Following formulae discovery is finally utilized to obtain the initial value of moving cell granting sequence vector
Step 3: determine moving cell number: utilize time domain subtraction, obtains moving cell number from sEMG, as cluster number.Detailed process is: the initial value first providing sequence vector from moving cell
in, find second largest peak value, remember that corresponding to this second largest peak value, the moment is
, secondly calculate
moment provides sequence, and computing formula is as follows:
Then sequence is provided
in find moment corresponding to K peak-peak, generally get
.Get the array sEMG value corresponding to K moment, phase adduction, divided by K, obtains the average in K moment.Finally in time domain, successively subtract K moment wave-average filtering value with array sEMG signal, stop until result is negative, record subtracts each other number of times M, this M value and moving cell number.
Step 4: be poised for battle column sEMG waveform clustering: according to described cluster number M, utilize minimum distance classifier, is poised for battle column sEMG waveform clustering.
Adopt Euclidean distance in minimum distance classifier, Euclidean distance formula is expressed as:
Wherein
the value of array sEMG corresponding sometime,
it is the meansigma methods of array sEMG value corresponding to K moment.
Step 5: calculate new moving cell and provide sequence vector: according to above-mentioned waveform clustering result, get the class containing maximum moment, calculate such moment waveform average, calculate new granting sequence vector.
The computing formula that new moving cell provides sequence is expressed as:
Wherein
it is N number of wave-average filtering value transposition.
Step 6: to all granting sequence classified finishings: repeat step 2--step 5, providing the moment can not extract until initial, and sEMG has extracted.Reject that repeat and irrational granting sequence vector, optimum results.Irrational granting sequence refers to the sequence that the granting moment is less than 15 milliseconds, needs to reject.
Claims (4)
1. an array sEMG signal automatic classifying method, is characterized in that comprising the following steps:
Step one: array sEMG Signal Pretreatment: to sEMG signal filtering, rejects interference;
Step 2: calculate and initially provide sequence vector: the dependency utilizing each channel signal of array sEMG signal, calculates the initial granting sequence vector of second largest peak value, as extraction initial value;
Step 3: determine moving cell number: utilize time domain subtraction, deducts the meansigma methods of waveform corresponding to K peak-peak from sEMG, obtains moving cell number, as cluster number;
Step 4: be poised for battle column sEMG waveform clustering: according to described cluster number, utilize minimum distance classifier, is poised for battle column sEMG waveform clustering;
Step 5: calculate new moving cell and provide sequence vector: according to described waveform clustering result, get the class containing maximum moment, calculate such moment waveform average, calculate new granting sequence vector;
Step 6: to all granting sequence classified finishings: repeat step 2-----step 5, rejects and repeats and irrational granting sequence vector, optimum results.
2. require a kind of described array sEMG signal automatic classifying method according to right 2, it is characterized in that: adopt the granting sequence in moment corresponding to second largest peak value as initial vector.
3. require a kind of described array sEMG signal automatic classifying method according to right 2, it is characterized in that: utilize K moment wave-average filtering value, by time domain subtraction determination cluster number.
4. require a kind of described array sEMG signal automatic classifying method according to right 3, it is characterized in that: in minimum distance classifier, adopt Euclidean distance, using sEMG waveform as class object, improve and decompose reliability.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510512579.4A CN104997508B (en) | 2015-08-19 | 2015-08-19 | Array sEMG signal automatic classifying methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510512579.4A CN104997508B (en) | 2015-08-19 | 2015-08-19 | Array sEMG signal automatic classifying methods |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104997508A true CN104997508A (en) | 2015-10-28 |
CN104997508B CN104997508B (en) | 2018-05-08 |
Family
ID=54370578
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510512579.4A Expired - Fee Related CN104997508B (en) | 2015-08-19 | 2015-08-19 | Array sEMG signal automatic classifying methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104997508B (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105326501A (en) * | 2015-12-10 | 2016-02-17 | 宁波工程学院 | Muscle disease monitoring method based on sEMG |
CN105342610A (en) * | 2015-12-10 | 2016-02-24 | 宁波工程学院 | Method for quickly evaluating depth of muscular movement unit |
CN105662336A (en) * | 2015-12-23 | 2016-06-15 | 黑龙江科技大学 | Signal de-noising processing method and apparatus |
CN105975917A (en) * | 2016-04-28 | 2016-09-28 | 宁波工程学院 | Array type surface electromyogram signal decomposition method facing high interference |
CN106264517A (en) * | 2016-09-30 | 2017-01-04 | 浙江大学 | A kind of method and system selecting electrocardio to measure position |
CN107526952A (en) * | 2016-06-22 | 2017-12-29 | 宁波工程学院 | Personal identification method based on multi-channel surface myoelectric signal |
CN108378848A (en) * | 2018-02-11 | 2018-08-10 | 宁波工程学院 | Muscular movement unitary space location estimation method |
CN108403114A (en) * | 2018-02-11 | 2018-08-17 | 宁波工程学院 | A kind of array Decomposition Surface EMG method towards constant force |
CN110720911A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit extraction method |
CN110720910A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit searching method based on correlation |
CN110720912A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit extraction method based on waveform correlation |
CN111419230A (en) * | 2020-04-17 | 2020-07-17 | 上海交通大学 | Surface electromyogram signal acquisition system for decoding motion unit |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101488189A (en) * | 2009-02-04 | 2009-07-22 | 天津大学 | Brain-electrical signal processing method based on isolated component automatic clustering process |
CN102961203A (en) * | 2012-12-10 | 2013-03-13 | 杭州电子科技大学 | Method for identifying surface electromyography (sEMG) on basis of empirical mode decomposition (EMD) sample entropy |
-
2015
- 2015-08-19 CN CN201510512579.4A patent/CN104997508B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101488189A (en) * | 2009-02-04 | 2009-07-22 | 天津大学 | Brain-electrical signal processing method based on isolated component automatic clustering process |
CN102961203A (en) * | 2012-12-10 | 2013-03-13 | 杭州电子科技大学 | Method for identifying surface electromyography (sEMG) on basis of empirical mode decomposition (EMD) sample entropy |
Non-Patent Citations (2)
Title |
---|
宁勇: "多通道表面肌电信号分解的研究", 《中国博士学位论文医药卫生科技辑》 * |
杨文元: "表面肌电信号的分解算法研究", 《计算机应用与软件》 * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105342610B (en) * | 2015-12-10 | 2017-12-22 | 宁波工程学院 | Muscular movement unit depth estimates fast method |
CN105342610A (en) * | 2015-12-10 | 2016-02-24 | 宁波工程学院 | Method for quickly evaluating depth of muscular movement unit |
CN105326501A (en) * | 2015-12-10 | 2016-02-17 | 宁波工程学院 | Muscle disease monitoring method based on sEMG |
CN105662336A (en) * | 2015-12-23 | 2016-06-15 | 黑龙江科技大学 | Signal de-noising processing method and apparatus |
CN105975917B (en) * | 2016-04-28 | 2019-06-21 | 宁波工程学院 | Towards strongly disturbing array Decomposition Surface EMG method |
CN105975917A (en) * | 2016-04-28 | 2016-09-28 | 宁波工程学院 | Array type surface electromyogram signal decomposition method facing high interference |
CN107526952B (en) * | 2016-06-22 | 2020-09-01 | 宁波工程学院 | Identity recognition method based on multi-channel surface electromyographic signals |
CN107526952A (en) * | 2016-06-22 | 2017-12-29 | 宁波工程学院 | Personal identification method based on multi-channel surface myoelectric signal |
CN106264517B (en) * | 2016-09-30 | 2019-05-14 | 浙江大学 | A kind of method and system selecting electrocardio measurement position |
CN106264517A (en) * | 2016-09-30 | 2017-01-04 | 浙江大学 | A kind of method and system selecting electrocardio to measure position |
CN108378848A (en) * | 2018-02-11 | 2018-08-10 | 宁波工程学院 | Muscular movement unitary space location estimation method |
CN108403114A (en) * | 2018-02-11 | 2018-08-17 | 宁波工程学院 | A kind of array Decomposition Surface EMG method towards constant force |
CN108403114B (en) * | 2018-02-11 | 2021-02-02 | 宁波工程学院 | Array type surface electromyographic signal decomposition method facing constant force |
CN110720911A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit extraction method |
CN110720912A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit extraction method based on waveform correlation |
CN110720910A (en) * | 2019-10-12 | 2020-01-24 | 宁波工程学院 | Muscle movement unit searching method based on correlation |
CN110720910B (en) * | 2019-10-12 | 2022-03-29 | 宁波工程学院 | Muscle movement unit searching method based on correlation |
CN110720912B (en) * | 2019-10-12 | 2022-04-26 | 宁波工程学院 | Muscle movement unit extraction method based on waveform correlation |
CN110720911B (en) * | 2019-10-12 | 2022-08-30 | 宁波工程学院 | Muscle movement unit extraction method |
CN111419230A (en) * | 2020-04-17 | 2020-07-17 | 上海交通大学 | Surface electromyogram signal acquisition system for decoding motion unit |
Also Published As
Publication number | Publication date |
---|---|
CN104997508B (en) | 2018-05-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104997508A (en) | Automatic decomposition method of array type sEMG (surface EMG) signal | |
Nazarpour et al. | Application of higher order statistics to surface electromyogram signal classification | |
CN102961203B (en) | Method for identifying surface electromyography (sEMG) on basis of empirical mode decomposition (EMD) sample entropy | |
CN104783787B (en) | A kind of J wave detecting methods based on neutral net | |
CN110059564B (en) | Feature extraction method based on power spectral density and cross-correlation entropy spectral density fusion | |
WO2016023349A1 (en) | Movement recognition method and apparatus based on surface electromyographic signal | |
CN109359619A (en) | A kind of high density surface EMG Signal Decomposition Based method based on convolution blind source separating | |
CN110163128A (en) | The Method of EEG signals classification of improved EMD algorithm combination wavelet package transforms and CSP algorithm | |
CN110321783A (en) | A kind of MEG spike detection method and system based on 1D convolutional neural networks | |
CN105956547A (en) | Decomposition method based on array surface electromyogram signal smoothing | |
CN107440716A (en) | Human body lower limbs athletic performance classification discrimination method based on single channel electromyographic signal | |
CN103761424A (en) | Electromyography signal noise reducing and aliasing removing method based on second-generation wavelets and ICA (independent component analysis) | |
CN104510468A (en) | Character extraction method and device of electroencephalogram | |
CN104571504A (en) | Online brain-machine interface method based on imaginary movement | |
CN105975917B (en) | Towards strongly disturbing array Decomposition Surface EMG method | |
Suhaimi et al. | Analysis of EMG-based muscles activity for stroke rehabilitation | |
CN107526952B (en) | Identity recognition method based on multi-channel surface electromyographic signals | |
Tariq et al. | Low SNR neural spike detection using scaled energy operators for implantable brain circuits | |
CN110647857A (en) | Gait recognition method and system | |
KR101446183B1 (en) | A Realtime EMG(Electromyograph) Pattern Recognition Method Employing Correlation Feature Analysis | |
Vigneshwari et al. | Analysis of finger movements using EEG signal | |
Rekhi et al. | Multi-class SVM classification of surface EMG signal for upper limb function | |
CN108836305A (en) | A kind of ECG feature extracting method of fusion Butterworth filtering and wavelet transformation | |
Tomczyński et al. | Hand gesture-based interface with multichannel sEMG band enabling unknown gesture discrimination | |
CN108573207A (en) | EMD and CSP merges most optimum wavelengths space filtering brain electrical feature extracting method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180508 Termination date: 20200819 |